26521 2009-06-18 15:35:47 -0700 Expose file size to chromium 2009-07-02 15:23:25 -0700 1 1 1 Unclassified WebKit Platform 528+ (Nightly build) PC Windows Vista RESOLVED FIXED P2 Normal --- 1 victorw victorw abarth darin dglazkov fishd playmobil tony oldest_to_newest 126790 0 victorw 2009-06-18 15:35:47 -0700 Need to implement getFileSize in WebCore/platform/chromium/FileSystemChromium.cpp. 126804 1 31515 victorw 2009-06-18 16:04:50 -0700 Created attachment 31515 Expose file size to chromium 126806 2 tony 2009-06-18 16:07:03 -0700 What is this method used for? I'm worried that an 0wned renderer can start reading to see if certain files exist or not. 126820 3 victorw 2009-06-18 17:14:50 -0700 (In reply to comment #2) > What is this method used for? I'm worried that an 0wned renderer can start > reading to see if certain files exist or not. > This is only allowed in uploading file. see bug: http://code.google.com/p/chromium/issues/detail?id=9102 Here is the Chromium patch that supports this: http://codereview.chromium.org/131082 126829 4 31515 tony 2009-06-18 17:34:50 -0700 Comment on attachment 31515 Expose file size to chromium I see, we validate the filename on the browser side. Seems like it would be nice if we could get the filesize information when the user picks a file and pass it through the IPC then (the FileChooser in the RenderFileUploadControl would get filename, size pairs). It would avoid the extra calls IPC calls back to the browser, but this seems like the easier change to make. 126836 5 victorw 2009-06-18 17:44:33 -0700 I was thinking about getting the filesize info and passing it through IPC with file name user selected, but the FileChooser allows selecting multiple files and more changes on the existing IPC. Discussed with Darin and think having a separate IPC to browser seems easier. (In reply to comment #4) > (From update of attachment 31515 [review]) > I see, we validate the filename on the browser side. > > Seems like it would be nice if we could get the filesize information when the > user picks a file and pass it through the IPC then (the FileChooser in the > RenderFileUploadControl would get filename, size pairs). It would avoid the > extra calls IPC calls back to the browser, but this seems like the easier > change to make. > 126860 6 31515 eric 2009-06-18 18:51:41 -0700 Comment on attachment 31515 Expose file size to chromium Looks fine to me. 126863 7 victorw 2009-06-18 19:22:23 -0700 (In reply to comment #6) > (From update of attachment 31515 [review]) > Looks fine to me. > could you hold landing this patch? would like to wait for the change in Chromium codebase to be reviewed. Will update this bug once the Chromium one is ready. 127397 8 fishd 2009-06-22 11:25:24 -0700 (In reply to comment #4) ... > Seems like it would be nice if we could get the filesize information when the > user picks a file and pass it through the IPC then (the FileChooser in the > RenderFileUploadControl would get filename, size pairs). It would avoid the > extra calls IPC calls back to the browser, but this seems like the easier > change to make. This seems like an interesting idea to me. It would mean changing FileChooser to hand back not just a list of file names but also the sizes of those files. One downside is that the file size returned to script would not be dynamic. That may actually be a deal breaker since I think it is normally valid for users to change the file to be uploaded before actually submitting the form. 127404 9 darin 2009-06-22 11:32:04 -0700 (In reply to comment #8) > I think it is normally valid for users to > change the file to be uploaded before actually submitting the form. Yes, it normally is. 128058 10 eric 2009-06-24 18:30:29 -0700 Can this be landed now? 128091 11 victorw 2009-06-24 21:03:42 -0700 Not yet. Still waiting for the patch in chromium. 128871 12 eric 2009-06-29 13:26:02 -0700 Looks like http://codereview.chromium.org/131082 was just approved by fishd. Let me know when this should go in. Any commiter can easily land this with: bugzilla-tool land-patches 26521 --no-build (--no-build because there is no way to build these chromium-only changes in WebKit). 128874 13 victorw 2009-06-29 13:37:56 -0700 Thanks Eric! Both patches are ready, but we need to sync the webkit patch with the Chromium change. Dimitri is the sheriff for Chromium webkit integration and he is going to land this one while he is doing the merging. (In reply to comment #12) > Looks like http://codereview.chromium.org/131082 was just approved by fishd. > Let me know when this should go in. > > Any commiter can easily land this with: > > bugzilla-tool land-patches 26521 --no-build > > (--no-build because there is no way to build these chromium-only changes in > WebKit). > 129596 14 dglazkov 2009-07-02 15:23:25 -0700 Landed as http://trac.webkit.org/changeset/45494. 31515 2009-06-18 16:04:50 -0700 2009-06-18 18:51:41 -0700 Expose file size to chromium webkit_getfilesize.patch text/plain 2144 victorw SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA0NDgzMikKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTYgQEAKKzIwMDktMDYtMTggIFZpY3RvciBXYW5nICA8dmljdG9yd0BjaHJvbWl1 bS5vcmc+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAg aHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTI2NTIxCisgICAgICAgIEV4 cG9zZSBmaWxlIHNpemUgdG8gY2hyb21pdW0KKworICAgICAgICBJbXBsZW1lbnQgZ2V0RmlsZVNp emUoKSBmb3IgQ2hyb21pdW0uCisKKyAgICAgICAgKiBwbGF0Zm9ybS9jaHJvbWl1bS9DaHJvbWl1 bUJyaWRnZS5oOgorICAgICAgICAqIHBsYXRmb3JtL2Nocm9taXVtL0ZpbGVTeXN0ZW1DaHJvbWl1 bS5jcHA6CisgICAgICAgIChXZWJDb3JlOjpnZXRGaWxlU2l6ZSk6CisKIDIwMDktMDYtMTggIFBl dGVyIEthc3RpbmcgIDxwa2FzdGluZ0Bnb29nbGUuY29tPgogCiAgICAgICAgIEZpeCBidWlsZCBi dXN0YWdlLgpJbmRleDogV2ViQ29yZS9wbGF0Zm9ybS9jaHJvbWl1bS9DaHJvbWl1bUJyaWRnZS5o Cj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT0KLS0tIFdlYkNvcmUvcGxhdGZvcm0vY2hyb21pdW0vQ2hyb21pdW1CcmlkZ2Uu aAkocmV2aXNpb24gNDQ4MzIpCisrKyBXZWJDb3JlL3BsYXRmb3JtL2Nocm9taXVtL0Nocm9taXVt QnJpZGdlLmgJKHdvcmtpbmcgY29weSkKQEAgLTgyLDYgKzgyLDkgQEAgbmFtZXNwYWNlIFdlYkNv cmUgewogICAgICAgICAvLyBETlMgLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLQogICAgICAgICBzdGF0aWMgdm9pZCBwcmVmZXRj aEROUyhjb25zdCBTdHJpbmcmIGhvc3RuYW1lKTsKIAorICAgICAgICAvLyBGaWxlIC0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLQor ICAgICAgICBzdGF0aWMgYm9vbCBnZXRGaWxlU2l6ZShjb25zdCBTdHJpbmcmIHBhdGgsIGxvbmcg bG9uZyYgcmVzdWx0KTsKKwogICAgICAgICAvLyBGb250IC0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLQogI2lmIFBMQVRGT1JNKFdJ Tl9PUykKICAgICAgICAgc3RhdGljIGJvb2wgZW5zdXJlRm9udExvYWRlZChIRk9OVCBmb250KTsK SW5kZXg6IFdlYkNvcmUvcGxhdGZvcm0vY2hyb21pdW0vRmlsZVN5c3RlbUNocm9taXVtLmNwcAo9 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09Ci0tLSBXZWJDb3JlL3BsYXRmb3JtL2Nocm9taXVtL0ZpbGVTeXN0ZW1DaHJvbWl1 bS5jcHAJKHJldmlzaW9uIDQ0ODMyKQorKysgV2ViQ29yZS9wbGF0Zm9ybS9jaHJvbWl1bS9GaWxl U3lzdGVtQ2hyb21pdW0uY3BwCSh3b3JraW5nIGNvcHkpCkBAIC0zMSw2ICszMSw3IEBACiAjaW5j bHVkZSAiY29uZmlnLmgiCiAjaW5jbHVkZSAiRmlsZVN5c3RlbS5oIgogCisjaW5jbHVkZSAiQ2hy b21pdW1CcmlkZ2UuaCIKICNpbmNsdWRlICJOb3RJbXBsZW1lbnRlZC5oIgogI2luY2x1ZGUgIlBs YXRmb3JtU3RyaW5nLmgiCiAKQEAgLTQ4LDEwICs0OSw5IEBAIGJvb2wgZGVsZXRlRW1wdHlEaXJl Y3RvcnkoY29uc3QgU3RyaW5nJikKICAgICByZXR1cm4gZmFsc2U7CiB9CiAKLWJvb2wgZ2V0Rmls ZVNpemUoY29uc3QgU3RyaW5nJiwgbG9uZyBsb25nJiByZXN1bHQpCitib29sIGdldEZpbGVTaXpl KGNvbnN0IFN0cmluZyYgcGF0aCwgbG9uZyBsb25nJiByZXN1bHQpCiB7Ci0gICAgbm90SW1wbGVt ZW50ZWQoKTsKLSAgICByZXR1cm4gZmFsc2U7CisgICAgcmV0dXJuIENocm9taXVtQnJpZGdlOjpn ZXRGaWxlU2l6ZShwYXRoLCByZXN1bHQpOwogfQogCiBib29sIGdldEZpbGVNb2RpZmljYXRpb25U aW1lKGNvbnN0IFN0cmluZyYsIHRpbWVfdCYgcmVzdWx0KQo=